Devitrification Dikes and Giant Spherulites Froni Klondyke, Arizona1

,N MINERALOGIST, VOL. 47, ]ULY-AUGUST, 1962

DEVITRIFICATION DIKES AND GIANT SPHERULITES FRONI KLONDYKE, ARIZONA1

Fner.Ir S. SrnroNs,U. S. GeologicalSuraey, Federol Center, Denaer,Colorad,o,

Aesrxlcr Vitrophyric rhyolite welded tufi near Klondyke, Arizona, contains thin dike-like sheets and spheroids as much as 6 feet across that have formed by devitrification of the original glassy rock. Chemical analyses show the devitrified tufi to have markedly more potash and less soda than its vitreous equivalent. Devitrification is believed to have been an autometamorphic process brought about by passage of potassium-rich fluids along fractures during weiding and cooling of the tuff. A similar chemical trend is noted in some devitrified rocks of Great Britain. fNrnonucrtor.t During the course of mapping the Klondyke quadrangle, Arizona, some striking devitrification features were noted in a rhyolite welded tuff. A cursory review of the literature reveals only scattered observa- tions of similar features, and it therefore seemsworthwhile to call atten- tion to the exceptionally well-exposedexamples at Klondyke.

GBor,ocrc SnrrrNc The Klondyke quadrangle is in Graham County, Lrizona, about 55 miles northeast of Tucson. The northeast quarter of the quadrangle is occupied by parts of the north-northwesterly-trendingSanta Teresa- Turnbull Mountains mass,the bulk of which is made up of sedimentary rocks of Paleozoicand Late Cretaceousage, plulonic igneousrocks of Precambrian and Tertiary age, and metamorphic rocks of Precambrian age. Along the west flank of the mountains is a thick sequenceof domi- nantly silicic volcanic rocks of possibleLate Cretaceousto early Tertiary age which either are faulted against or rest unconformably on the older rocks. The rock in which the devitrification features were noted is ex- posedalong the east side of a gully in the center of sec. 18, T.6 S., R. 20 E., about 2,000feet eastof the road to Imperial lVlountainand 5 miles north of Klondyke. In this area the volcanic rocks are silicic lavas, flow breccias, tufis, and welded and partly welded tuffs that strike N. 25"-30"W. and dip 55'-60'SW. One unit of the sequenceis a lens of black vitrophyre as much as 40 feet thick that is exposedfor 900 to 1,000feet along the strike (Fig. 1). The lens appears to pinch out to the northwest and is covered by al- luvium at its southeast end. The vitrophyre is a hard glassy rock with subconchoidal to somewhat hackly fracture and contains scattered

1 Publication authorized by the Director, U. S. Geological Survey. 87r 872 FRANK S. SIMOTS

Frc. 1. Lens of vitrophyric rhyolite welded tuff (B) dipping 55o-60o to the right and somewhat toward the observer. The lens is about 40 feet thick. It is underlain by pale red partly welded tufi (A) and overlain by grayish red welded tufi (C). The devitrification "dikes" are best exposed on the steep slope below the highest outcrop; the giant spherulites crop out along the entire slope and seem to be concentrated near the top of the vitrophyric welded tuff just beneath unit C. Arrows point to outcrops of several giant spherulites. millimeter-sized phenocrysts of feldspar and sparse biotite and quartz It gradesdownward through 15 20 feet of mottled black and red partly vitrophyric welded tuff to a sharp contact with pale redl partly welded rhyolite tufi 10 30 feet thick. The vitrophyre lens is overlain by grayish red solidly welded and almost completely devitrified rhyolite tuff of unknown but appreciable thickness. All these rocks are believed to be parts of a single tuff deposit, the lithologic variations being due mainly to differencesin degree of welding and devitrification. Under the microscope the vitrophyre consists of densely welded pale brown glass shards enclosing euhedral to subhedral grains of oligoclase, sanidine, and minor brown biotite and quartz, as well as sparsepumice lapilli and lithic fragments (Fig. 2).

DBvrrnrrrcarroN DrKES The vitrophyric welded tufi is cut by numerous steep-dipping dike- like bodies of moderate brown devitrified rock that contrast strongly with the black vitrophyre (Fig. 3) and stand out in slight relief from the vitrophyre upon weathering. These "dikes" are confined to the vitrophyre, although at least one "dike" structure is representedin the under-

I This and succeeding colors are from the Rock Color Chafi, Nati.onal ResearchCouncitr, 1948. DEVITRIFICATION DIKES 873

Frc. 2. Photomicrograph of vitrophyric rhyolite welded tufi. {-plagioclase; b-biotite; l-lithic fragment; p-pumice; rest of field is pale-brown glass shards. The vitroclastic texture is clearly preserved but the shards are densely welded. Plane polarized light' X50.

Frc. 3. Light colored devitrification "dikes" in black vitrophyric welded tufi. Scale is 7] inches long, widest "dike" is about 14 inches wide. AII these "dikes" pinch out uphill just above field of photograph. The vitrophyre between the two largest "dikes" is cut by several smaller "dikes" not clearly visible in the photograph 874 FRANK S. S/MOTS

Frc. 4. Light-colored narrow devitrifica- Frc. 5. Photomicrograph of irregular tion "dikes" in black vitrophyric welded contact between vitrophyre (v) and "dike" tufi. Scale is 7] inches long. Note the thin of devitrified rock (d). Textures on both white median layer of quartz (q) in both sides of the contact are essentially identical "dikes," particularly the right-hand "dike." except where the original texture has been Left-hand "dike" continues uphill for 15-20 destroyed by growth of incomplete spheru- feet above field of photograph; both "dikes" lites, as along the upper right-hand segment pinch out just below field. of the contact. Plane polarized light, X33. lying partly welded tufi by a simple fracture. The "dikes" are rather irregular in width, ranging from a fraction of an inch to about 3 feet across;most are an inch or less wide. Near the southeastend of the vitrophyre lens, numerous ramifying "dikes" divide the vitrophyre into a mosaic of irregular blocks. All the "dikes" are divided symmetrically along their centers by a thin layer of qlrartz. The thickness of the quartz layer appears to be roughly proportional to the width of the enclosing "dike" but is less than 2 millimeters regardlessof "dike" width. Contacts between vitrophyre and "dikes" may be irregular with many sharp bends,re-entrants, etc., as shown in Fig. 3, or may instead be essentially straight for appreciable distances, as in Fig. 4. The contacts invariably are sharp, and no transition from stony to vitrophyric is de- tectable with a hand lens. Within the "dikes" themselves,however, the rock flanking the quartz core is slightly darker than that in contact DEVITRIFICATION DIKES 875

with the vitrophyre, the color changing gradually outward from moderate brown to light brown. The core rock also is more resistant to weatheringthan is that along the contacts,which on weatheredsurfaces commonly are marked by a shallowtrench resultingfrom the disintegra- tion of rock along both sidesof the contact. On a microscopicscale the contacts between,,dike,' and vitrophyre are exceedingly irregular, with narrow tongues of one rock penetrating the other (Fig. 5). The contactsare, however,quite sharp, and provide almost no evidenceof any transition betweenthe two rocks.Locally, the original vitroclastic texture has been obliterated by the formation of incomplete spherulites, but for the most part is clearly preserved.in the devitrified rock. only a few tiny islands of glassy rock have survived devitrification. Within the devitrified rock, none of the phenocrystic minerals appear to have been altered at all during devitrification. The quartz seamsalong the centersof the ,,dikes',consist of a core of granular qutrtz flanked by narrow irregular selvagesmade up of tiny rhomb-shapedgrains of adularia (?) in a fine-grainedmatrix that may be cristobaliteor tridymite (Fig. 6); in generalthe mineralsof the selvages are too fine grained to be identified positively in thin section. The straight sharply definedwalls of the quartz seamsand the completelack of any evidencesuggesting replacement of the host rock indicate that the seams must be true fissure fillings. The sequenceof development appears to have been: 1) fracturing of the welded tuff to produce narrow open fissures;2) depositionof a little very fine-grainedadularia (?) and cristobalite (?) or tridymite (?) along the walls of the fissuresland 3) complete filling of the fissures by quartz. Some of the quaftz was de- positedearly in the third stage,as witnessthe two grainswith hexagonal outlines near the left edge of Fig. 6, which must have been deposited prior to the final filling of the fissure. The minerals formed during devitrification are too fine grained to be identified by optical methods. A sample of devitrified tufi carefully hand picked to remove phenocrystic minerals was examined with the r-ray spectrometerby Theodore Botinelly of the U. S. GeologicalSurvey, who reports plagioclase feldspar, qvartz, cristobalite, orthoclase and mica. Chemical analyses of a devitrification ,,dike,, and of its immediately adjacent vitrophyric wall rock are given in Table 1, columns I and,2; these samples must represent essentially identical original rock. The principal chemical change accompanying devitrification obviously has been addition of potash and removal of soda; the ratio K2O:Na2O is about 1.4 in the glassyrock and 7.7 inthedevitrified, rock. The devitrified equivalent also is slightly more silicic and contains about 2 per cent less 876 FRANK S. S/MOlilS

g:gl;9 9€l33ES ^i'.i ';+ iiqqK qq!qiEe dr ioHrN

€ rD O\OOO+6.6+ a D+ \oa rt'i-' ri - J

N Nor@@ 91q. o r

v N

@ d Nts OTNN @ cO € rOO r NO\OO + !.1 4 eN' O$

Fr r n*n-OOd\OrrN€'6i '<. O <. - J <, Fr N 6 r oO +1 4 c) + qr N N N d r S r c' F : J o o o o c \c @ id = - 6 cj -l tl 'Jr-Jci o. + 'qiGi -;'.iJ o.

9ES3E5RS?iG338833e9'd '*;.+ 'q' J -;-o\ o

.l 2 € b ddq*q'AaAnd>>AZVEEAaUr..TFA ? ?? ?"^{""

Fic. 6. Photomicrograph of median quartz seam in devitrification "dike." Wall rock is devitrified welded tuff. The quartz core, about 0.5 mm wide, is flanked by narrow irregular selvages composed of rhomb-shaped grains of adularia (?) in an irresolvable matrix that may be cristobalite or tridymite. Plane polarized light, X45.

1. Fresh vitroph)'ric rhyolite welded tuff, sec.18, T. 6 S., R. 20 E., Klondyke quadrangle, Arizona. Analyst: Dorothy F. Powers, U. S. Geological Survey. Field No. 59K-S-63a; lab. no. F2676. 2. Same rock as (1), devitrified. Analyst, Dorothy F. Powers. Field no. 59K-S-63b; lab. no. F2677. 3. Average alkali rhyolite plus rhyolite obsidian. Nockolds, 1954,Table 1, col. IV. 4. Glassy leidleite (:dacite) rvith "sheath and core" structure, Mull, Scotland. Ander- son and Radley, 1915, p. 212, col. Ia. 5. Stony (devitrified) leidleite, same locality as (4). Anderson and Radley, 1915, col. Ib. 6. Pitchstone porphyry from Cir Mhor dike, Arran, Scotland. Judd, 1893, p. 545, column III. 7. "Quartz-felsite" derived from (6). Judd, 1895, col. IL 8. Glassy leidleite, Mull, Scotland. Anderson and Radley, 1915, col. IIa. 9. Stony leidleite, same locality as (8). Anderson and Radley, 1915, col. IIb. 10. Glassy leidleite, Mull, Scotland. Anderson and Radley, 1915, col. IIIa. 11. Stony leidleite, same locality as (10). Anderson and Radley, 1915. col. IIIb. 12. Black obsidian flow, Colle de la Motte, Esterel, France. Terzaghi, 1948, Table 1, col.11. 13. Devitrified obsidian "a few hundred feet" from (12). Terzaghi, 1948, col. 7 . 14. Perlite, Wenatchee, Washington. Coombs, 1952, Table p. 202, col. 2. 15. Spherulitic breccia, presumably derived from (14). Coombs, 1952,Table p.202, col.3. 878 FRANK S. S/MOTS

Frc. 7. Giant spherulite in vitrophyric welded tufi. Scale is 7! inches long, spherulite is about 3 feet across. total water than the fresh rock; however, the weakly perlitic structure of the glasssuggests that at least someof the HzO+is not pristine, and if the original water content be taken as, say, 0.4 per cent, following Ross and Smith (1955, p. 1081-1082),and the analysesrecalculated, then the silica percentagesbecome nearly equal. Semiquantitativespectrographic analyses by Nancy M. Conklin, U. S. GeologicalSurvey, of 56 minor elementsin the two rocks show only the following changes(data arein per cent); all are of one order of magnitude, and none is believed to be significant. Element Vitrophyre Devitrified equivalent Ba .03 .07 Cr .00015 .0003 Ga .0015 .0007 Sr .015 .03 Yb .0003 .0007

GreNr Spunnurrrns The vitrophyric weldedtuff also containsnumerous spheroidal masses of devitrified rock identical lithologically with the devitrification "dikes." Field evidence,although inconciusive,suggests that the spheroidsare concentratedin the upper part of the vitrophyre lens. A number are clustered near the northwest end of the lens (Fig. 1). The spheroids range in diameter from a few inchesto 6 feet or more, and averageper- haps 3 feet across(Fig. 7). They consist of a core of moderate brown DEVITRIFICATION DIKES densestony material encasedby a thin crust, a few inchesthick at most, of light-brown somewhat crumbly rock. Most spheroidsobserved were divided roughly in half by a thin qtartz seam. One Iarge spheroid 5-6 feet across passesuphill abruptly into a narrow devitrification "dike," and it seemslikely that otherscould be shown to do so if exposureswere more complete. Microscopically,the rock composingthe spheroidsis identical with that of the "dikes." Although the spheroidsare demonstrably not the result of the simple spherulitic crystallization responsible for the com- mon spherulites of glassy volcanic rocks, but rather are compound- made up of a great many spherulitic, radiolitic, or axiolitic aggregates- neverthelesstheir origin and external form would seem to permit use of the name spherulite.

RBr-erBo Fparunrs Fnolr Ornpn Locettups The pitchstonesills and sheetsof Mull (Andersonand Radley, 1915) exhibit phenomena very similar to the devitrification "dikes" at KIon- dyke. These sills are commonly composedof Ieidleite (a rock near dacite in chemicalcomposition). A typical sill is 11-12 feet thick and consists of a core 5 feet or so thick of glassy rock or pitchstone and margins per- haps 3 feet thick of stony devitrified rock. The transition between pitch- stoneand stony rock is abrupt but the stony portion doesnot resemblea chilled margin. The pitchstone may be split by stony partings. A feature of these rocks, first describedfrom the Eskdale pitchstone by Geikie (1880) as "sheath and core" structure is discussedin some detail by Andersonand Radley (1915),who say (p.210):

"The stony [devitrified] base and top of the intrusion send ofi arms, or, more accurately, narrow sheets of sirnilar material, which traverse the glassy portion in a branching and sinuous manner, completely dividing it into irregular rounded cores, which are not in visible contact. In many cases the sheaths of crystalline matter, which vary from a quarter of an inch to 3 inches in thickness, show a median joint or suture-line, along which terminate numerous minute transversejoints."

Along one median suture noted in thin section, the first mineral to form was quartz, followed by "a thin layer of some mineral having a very low refractiveindex;then by further qrartz, and in placeslimonite" (Ander- son and Radley, 1915,p. 213-214). The Cir Mhor dike of Arran in westernScotland (Judd, 1893,p. 543- 551) is a compositedike 12-30 feet wide with a center of highly silicic rock and selvagesof andesite. The silicic dike in turn has a core of pitchstone porphyry flanked by a devitrified equivalent of the pitchstone called "quartz-felsite" by Judd. No cross-cutting devitrification features are mentioned,but the stony selvagesotherwise resemble closely 880 FRANK S SlMOlilS the devitrification "dikes" and are considered by Judd to be results of primary devitrification, that is, devitrification taking place during con- solidationof the dike (Bonney and Parkinson,1903). Terzaghi (1948,p. 24) noted that devitrification is limited to the im- mediatevicinity of cracksin an obsidianof the Esterel,France. Spherulites as large as those from Klondyke apparently are rather rare; at least few are mentionedin the literature examined.Cross (1891; 1896,p. 298-299,Pl. 28) has describedgiant compound spherulitesin pitchstone at the baseof a rhyolite flow at Silver Clifi, Colorado; here many of the spherulitesare 5 feet in diameter and some are more than 10 feet across.The large spherulitesare embeddedin a "white mixture of opaline silica and kaolin resulting from decompositionof the pitchstone . . ". According to Iddings (1909, p. 239-240, Fig. 77), these spherulites"appear to have formed by the rapid growth of rays of plu- mose spherulitic aggregatesoutward from a central core and the final radial growth of a denser layer forming an outer spherical shell. The whole of the inclosedrock, or magma, is not necessarilycrystallized in radial arrangement.It may even be glassyin part." Fuller (1931,p. 64) noted some spherulitesas much as 3 feet in diameter in the "upper laminated rhyolite" of SteensMountain, Oregon.

Crrnlrrc-qr,C u.qNcr s AccoupewvrNG D EvrrRrFrcATroN Glass is generally consideredto be an undercooledhighly viscous (rigid) liquid (NIorey, 1954, p. 28), although some investigatorshave referredto glassas an amorphoussolid in recognitionof certain changes in physical propertiessaid to occur upon cooling when the glassattains a particular viscosity (for instance, Ilawkes, L930). X-ray difiraction studies of various glassesby Warren and others (summarizedin Stan- worth, 1950,p. 16-31)have shown the structure of glassto resemblethat of a very distorted crystal lattice. Devitrification, or crystallization, of natural or artificial glassesmay take place, provided that conditions permit the formation of stable nuclei, simply becausean undercooled Iiquid or melt is thermodynamically unstable, and therefore may occur within a closed system; many spherulites in volcanic glassesmust have formed under such conditions. Wholesale devitrification of natural glasseshas been attributed to heat, pressure,circulating fluids, "power of spontaneouscrystallization" or a combination of thesefactors (Harker, 1909,p. 226;Bonney, 1885,p. 82-95). Devitrification of the Klondyke vitrophyric tuff, however, is clearly related to fractures and must have been causedby fluids passingalong the fractures. Devitrification cannot be attributed solelyto heating by thesefluids of the glassalong the fractures, becausethe chemicalanalyses show that a marked compositional DEVITRIFICATION DIKES 881 changehas occurred.Nor is it likely that the chemical changesnoted could be a result of weatheringor of the circulation of ground water. It is possiblethat devitrification was causedby later hydrothermal solu- tions or gasesfrom extraneous sources, but the nearby rocks are un- altered, the "dikes" are completely different from anything I have seen or read about in hydrothermally altered volcanic rocks, and the ap- parent restrictionof the "dikes" to the vitrophyre suggestsa local source for the fluids responsible for the devitrification. The most probable source of the added potassium is the welded tufi itself or the partly welded tuff underlying it. Presumably therefore the potassium-rich fluids were erupted and transported along with the tuff, were trapped subsequentlywithin it at the time of deposition,and later escapedalong contraction cracks formed during cooling of the vitrophyre and now marked by the devitrification "dikes." Welding, fracturing, expelling of interstitial potassium-rich fluid, and devitrification all may have occurred at about the sametime. Such a seriesof eventsmight seemrather improbable, but apparently these devitrification features are unusual enough, at least in the Klondyke region, to have required uncommon circumstances; similar vitrophyric welded tuffs crop out over several square miles within the quadrangle, and many miles of contact are well exposed,but at no other place have thesefeatures been observed. Paired analysesof fresh glassand its devitrified equivalent seem to be scarcein the literature, but a few are includedin Table 1. A trend in the KzO:NazO ratio similar to although less pronouncedthan that in the Klondyke rocks is evident in analysis pairs 4-5, (ratio increasesfrom 0.5 to 1.4), 6-7 (ratio increasesfrom 1.0 to 1.3), 12-13(ratio increases from 0.8 to 2.9),and 14-15 (ratio increasesfrom 0.3 to 1.4).The possibil- ity of minor original differencesin these pairs of rocks cannot be excluded as confidently as for the Klondyke pair, although Judd (1893) states with assurancethat (7) differsfrom (6) (this paper,Table 1) only in that the groundmassis devitrified. Anderson and Radley (1915, p. 2I0-2lI) attribute the formation of the "sheaths" of "sheath and core" structure in their leidleite sills to escapeof some volatile substancefrom the devitrified margins and along joints, and present three sets of paired analyses (Table 1, columns 4-5 and 8-11) showing the devitrified rock contains about 1 per cent less water above 105' C. than the glassy rock. However, their one pair of complete analyses (Table 1, columns 4-5) shows other changessuggest- ing that more than simple dehydration was involved in devitrification; specifically, that potassium was added and sodium subtracted. Terzaghi (1935, p. 378-379) has presented a variation diagram for 9 volcanic glassesand 4 devitrified volcanic glasses,all containing more 882 FRANK S. SlMOlrS than 72 per cent silica. The diagram shows the devitrified glassescontain somewhat less soda and distinctly more potash, as well as more water, than the non-devitrified glasses.Although no pairs of analyses are given, a general trend similar to that of the Klondyke rocks is ap- parent. Replacement of soda by potash accompanying devitrification was noted by Fenner (1934, p. 240-2431'1936, especially p. 262-277) in specimensof dacite pitchstone from a borehole in the Upper Basin of Yellowstone National Park. A pitchstone from a depth of 260 |eet (specimeny.P.279) contained 3.97 per cent NazO and 2.75 per cent K2O, whereas a slightly devitrified but otherwise similar rock from a depth of 225 leet (specimeny.P.267) had 3.38 per cent NazO and 3.59 per cent KrO; the ratio KzO: NazOincreased from 0.69 in the glassyrock to 1.06in the partly devitrified rock. Other rocks from depths of.293to 400 feet show still higher ratios of potash to soda and also a notable in- creasein silica, but their original similarity to the pitchstone is less as- suredl some show perlitic structures but others may not have been highly glassy rocks. Nfost non-devitrified natural glassesare young geologically, the great majority being of Cenozoicage (Harker, 1909,p. 225-226;Tyrrell, 1926, p. 82).The oldestone reported to date from North America seemsto be a vitrophyric rhyolitic(?) welded tuff of late Late Cretaceous(Montana) age from near Wolf Creek, Montana (Barksdale,1951). G. D. Robinson (personalcommunication, 1961) reports glassyrocks of about the same age from the Three Forks quadrangle, Montana. I{owever, some are reported to be much older. Glass of possible early Paleozoicor Precambrian age is reported from King Island, Tasmania by Scott (1951);glassy volcanic rocks are de- scribed from the Ordovician Borrowdale volcanic seriesof Great Britain by Mitchell (1929, p. 20; 1934, p. a2Q and Hartley (1925, p. 207-208) and from Ordovician rocks near Llangynog, Wales by Cantrill and Thomas (1906, p. 242 and Pl. 24, Fig. 2) ; numerous occurrencesof glass have been noted in andesite.basalt. and dacite of Old Red Sandstone (Devonian) age in Great Britain (Teall, 1883, p. 105,254-255; ltdd, i'n Durham, 1886,p. 427-429; Flett, 1897,p. 291; Geikie, 1897, v. 1, p. 274 Jowett,1913,p. 475; Balsillie,1918, p.349; Pringle, 1948,p.47; Macgregorand Nlacgregor,1948, Table oppositep.25;Harker, 1954,p. 166-167;Harry, 1956,p. 50); and glasshas beenidentified in andesitesof Carboniferousor Permian agein Great Britain (Geikie, 1897, v. 2, p. 45, 57; Pringle, 1948,p.65; Macgregorand Macgregor,1948, p.54-55). Many of these glassesare reported in the older literature, and clear DEVITRIFICATION DIKES 883 evidencethat the glasshas not devitrified is commonly not given, but the existenceof glassat least as old as Carboniferousseems well estab- lished (see,for instance, Judd, in Durham, 1886,p. 427-429). The composition and refractive index of the Tasmanian material are so unlike those of any ordinary volcanic glass that identification of this material as a true voicanicglass seems at least questionable. .Theseoccurrences of geologically old glasses,some of which are in de- formed rocks that must have beensubjected to temperaturesabove those prevailing at the surfaceat the time of extrusion,appear to be rare, but they suggestthat natural glassesare inherentty highly stable,that they seldom if ever devitrify solely as a result of aging, and that very appreciable heat or pressure may be required to devitrify them. Various explanations have been proposed for the rare and apparentiy anomalous old glasses,but none seemsgenerally valid. For instance,Hawkes (1930) points out that some of them have as much as 10 per cent water and suggeststhat a high water content may inhibit devitrification, but other old glassescontain only ordinary amounts of waterl furthermore, the importance of water in the devitrification of natural glassat moderate temperatures is strongly emphasizedby Marshall (1961). Chemical compositiondoes not appear to be a controlling factor, as the Paleozoic glassesfrom England and Scotland range from rhyolitic to basaltic in composition. CoNcrusroNs Vitrophyric rhyolite welded tuff at Klondyke, Ariz., has devitrified along fractures, and large spherulites have formed in the upper part of the tuff. The devitrified rock is shown to contain much more potassium and markedly less sodium than its glassy equivalent. Devitrification was an autometamorphic processbrought about by passagealong fractures of potassium-rich fluids believed to have been trapped within or below the welded tuff at the time of emplacement. Devitrification of many other natural glassesmay have had a similar deutericorigin, and indeed some such processmay be the only generally operative causeof devitrification; the occurrenceof geologicallyold non-devitrified glassessuggests that natural glasseshave a high inherent stability and, once having es- caped autometamorphic devitrification, may persist for a geologic era or longer. AcrNowrBocMENTS

I wish to thank ProfessorAaron Waters of The Johns Hopkins Univer- sity and my colleaguesRoy Bailey, James Ratte, and Thomas Steven for their constructivecriticism. 884 ttRANK S. SIMOIrS

Rrlrnrxcrs

ANonnsoN, E. M. aNn E. G. Reor,Bv (1915), The pitchstones of Mull and their genesis' With notes on the rock-species leidleite and inninmorite by H H. Thomas ancl E' B' Bailey. Geol. Soc.London Quart. J0w.,71,205-217. Bersrr-r,rr, D., (1918), Note on a hypersthene andesite from Pitcullo, Fifeshire. GeoI.Mag', 5, 34G350. Bemso.lr,e, Jur,r.rn D. (1951), Cretaceous glassy welded tuffs-Lewis and Clark County, Montana. Am. J ow. Sci., 249, 439-443. BoNNEy, T. G. (1885), Anniversary address of the president. GeoI. Soc. Lond'on Quort' Jour. Proe.,41 37-96. --- glassy ANDJ. P,mnNsom (1903), On primary and secondarydevitrification in igneous rocks. Geol. Soc.London Quart. four.,59, 429-444. CeNtnrr-r.,'I'. C. aNo H. H. Tnorrlas (1906), On the igneous and associatedsedimentary rocks of Llangynog (Caermarthenshie). Geot. Soc.London Quart. Jour',62,223-252. Coouns, H. A. (1952), Spheruiitic breccias in a dome near Wenatchee, Washington' '42' M ineral'., 37, 197-206. Cnoss, WmruAN (1891), Constitution and origin of spherulites in acid eruptive rocks. P hil,os. Soc. W oshington, BuIl., ll, 41l-M4. -- (1896), Geology of Silver Clifi and the Rosita Hills, Colorado. U. S' Geol'.Surwy 17th Ann. Rept., pt.2,263-403. Dunnnu, Jauns (1886),Volcanic rocks of the north-east of Fife, with an appendix by J. W' Jtd.d..Geol. Soe . Lond'onQuart. Iour.,42,418434. l-ENNER,C. N. (1934),Hydrothermal metamorphism in geyser-basinsof Yellowstone Park, as shown by deep drllling Am. Geophys.Union Trans.,15,240-243. --- (1936), Bore-hole investigations in Yellowstone Patk' Jour. GeoI,44,225-315' !'r.rrr, J. S. (1897), A hypersthene andesite from Dumyat (Ochils)' Ed'inburgh GeoI- Soc. Trans.,7, 290-297. Fulr,rn, R. E. (1931), The geomorphoiogy and volcanic sequenceof Steens Mountain in southeastern Oregon. Wasld,ngton Unizt. Publ" Geol'.,3, l-I3O. Grrrrr, Ancnmar,n (1880), The "pitchstone" (vitreous basalt) of Eskdale. Royal Phys- Soc.Edinburg Proe.,5, p. 219 lquoted by Andersonand Radley, 1915,p' 214)- --- (1897), The Ancient VolcanoesoJ Great Britoin. London, Macmillan and Co. Ltd', 2v. Ha.nxrn, Ar,rnnn (1909),The Nataral ll'istory oJIgneous Rochs.'London,Methuen and Co' -- (1954), Petrol'ogyJor Students,8th ed : Cambridge University Press' H.Lnnv, W. T. (1956),The OId Red Sandstonelavas of the western Sidlaw Hills, Perthshire. GeoI.Mag.,93, 43-56. Henrrrv, I . I . 0925) , The succession and structure of the Borrorvdale volcanic series as developedin the area lying between the lakes of Grasmere,windermere, and coniston. Geol.A ssoe.Lonilon Proc., 36, 203-226. HAwKES, Lrorqanr (1930), On rock glass, and the solid and liquid states. Geol Mog.,67, t7-24. IortNcs, Josmn P. (1909),I gneousRocks, v.1. John Wiley & Sons,New York' JowErr, Ar,snnr (1913), The volcanic rocks of the Forfarshire coast and the associated sediments. Geol..Soc. London Quart. Iour.,691 459-483' 49, Junn, J. W. (1893), On composite dykes in Arran. Geotr.Soc. Londnn Quart. f our', 536-565. 'Ihe Mlccnncot, M. aNo A. G., Maccnncon (1948), British regional geology. Midland Valley of Scotland. Great Bri.tai.n Geol,.Surtey. DEVITRIFICATION DIKES 885

Mansrl.tr.r,, R. R. (1961), Devitrification of natural glass. GeoI. Soc. Am. BuIl., 72, 1493-1520. Mrtcnnll, G. H. (1929), The succession and structure of the Borrowdale volcanic series in Troutbeck, Kentmere, and the western part of Long Sleddale (Westmorland) Geol'.Soc. Lonilon Quart. Iou.r.,85,9-44. (1934),The Borrowdale volcanic seriesand associatedrocks in the country between Long Sleddale and Shap. Geol.Soc. London Quart. Jour.,9O,4l8-M4. Monnv, G. W. (1954), The Properties of Glass,2nd ed. Reinhold, New York' Nocnor-rs, S. R. (1954), Average chemical compositions of some igneous tocks Geol'.Soe . Am. Bul,I.. 65. 1007-1032. PnrNcr,n, J. (1948), British Regional Geology. The South of Scotland. Great Britain GeoI- Suraey. Ross, C. S. ,q.NrR. L. Surtn (1955), Water and other volatiles in volcanic glasses.Am. Mineral,., 40, 1071-1089. Scorr, Bnnvr. (1951), The petrology of the volcanic rocks of southeast King Island, Tasmania. Royal Soc. Tasmania Papers Proc.1950, 113-136. SraNwonur, J. E. (1950), Physi,cal,Properties oJ Glass. Oxford Univ. Press, London. Tnrrr-, J. J. H. (1883), Notes on the Cheviot andesites and porphyrites. GeoI.Mog.,lO' 100-108. 145-152. 252-262. Tanzecur, R. D. (1935), The origin of potash-rich rocks. Am. Jour. Sci.,229,369-380. --- (1948), Potash-rich rocks of the Esterel, France. Am. Mineratr.33' 19-30. Tvnnrrr, G. W. (1926), The Prineiples oJ Petrol'ogy:E. P. Dutton and Co , New York.

Monuscript receioeil,Noaember 28, 1961.